hnguyen_1
14 years agoOccasional Contributor
Think Time, Master/Slave, etc.
Hi,
1) I was reading up on more load testing information and was told that "If an agent computer has more than 75% CPU utilization, or has less than 10% of physical memory available, it is overloaded." Is this correct? So if I want to simulate 3000 concurrent users, I would need to set up a master/slave configuration for load testing using the network suite? Usually, how many virtual users should be ran from 1 client computer, under what specifications?
2) I read http://msdn.microsoft.com/en-us/magazine/cc188783.aspx#S3, but I still don't understand how 1 virtual user translates to 10 - 15 real users. For example, I have a task for export PDF. I assign that to 1 virtual user, if I run that task for 15 virtual users, that will still be 15 concurrent export PDF at the same point in time, how does it translate to 10 * 15 real user performing the export PDF? So, according to that article, think time should not be present in load tests? How do I verify that my load test simulate real user interaction with the system besides just taking their word for it? Also, I've been reading up on how to create a load test scenario but I still don't know how to apply it to my project. Will someone from SmartBear be able to help me with that?
3) From http://smartbear.com/forums/forum/post/?mode=singleThread&thread=07c628a1-1344-4007-8542-b9f17162c15f, I thought that random think time would be introduced into TestComplete 8.1? How will that affect the reproducibility of the test? Is it better to use start delay if I want the results to be reproducible?
1) I was reading up on more load testing information and was told that "If an agent computer has more than 75% CPU utilization, or has less than 10% of physical memory available, it is overloaded." Is this correct? So if I want to simulate 3000 concurrent users, I would need to set up a master/slave configuration for load testing using the network suite? Usually, how many virtual users should be ran from 1 client computer, under what specifications?
2) I read http://msdn.microsoft.com/en-us/magazine/cc188783.aspx#S3, but I still don't understand how 1 virtual user translates to 10 - 15 real users. For example, I have a task for export PDF. I assign that to 1 virtual user, if I run that task for 15 virtual users, that will still be 15 concurrent export PDF at the same point in time, how does it translate to 10 * 15 real user performing the export PDF? So, according to that article, think time should not be present in load tests? How do I verify that my load test simulate real user interaction with the system besides just taking their word for it? Also, I've been reading up on how to create a load test scenario but I still don't know how to apply it to my project. Will someone from SmartBear be able to help me with that?
3) From http://smartbear.com/forums/forum/post/?mode=singleThread&thread=07c628a1-1344-4007-8542-b9f17162c15f, I thought that random think time would be introduced into TestComplete 8.1? How will that affect the reproducibility of the test? Is it better to use start delay if I want the results to be reproducible?